Dynatrace AssistDynatrace Assist: Ask, analyze, and act with Dynatrace Intelligence.
AppDynatrace Assist is your natural‑language gateway into Dynatrace Intelligence, the agentic operations system that brings the full platform together for you. Instead of switching between dashboards, queries, apps, and other tools, you simply ask a question and Dynatrace Intelligence responds accordingly, using Dynatrace’s generative, causal, and predictive AI capabilities as well as all agentic tools provided by Dynatrace MCP Server.
Like a true team member, Assist draws context from Grail, maps relationships through Smartscape, and collaborates with Dynatrace agents to deliver grounded, precise answers. Use the insights from Dynatrace experts with the product - from revealing insights, identifying patterns, troubleshooting, to best practices - all backed on domain expertise.
As the conversation continues, you naturally move from insights and understanding to exploration and action. Ask. Analyze. Act.
Assist ships ...
Use Assist for use cases such as:
Investigate incidents: ask Dynatrace and let Dynatrace Intelligence reason through a problem, using all available tools and skills.
Analyze your data, get insights, and recommended remediation steps.
Explain Kubernetes warning signals, Logs, Problems, threat detection findings, or database insights integrated right into the respective app
Interact with all available Dynatrace foundational agents for deeper analysis when needed
In-context assistance to help with onboarding, Dynatrace usage, and configuration topics
and much more coming in the next weeks and months.
To get started with this app, you need to enable
for your environment and assign the necessary permissions to the relevant user groups via a policy. See our documentation for further details.
Once this is done, users can select the Assist entry (below the global search) in the dock and ask Dynatrace Assist their desired question, or select one of the example prompts.
ℹ️ Curious to understand how Dynatrace Intelligence generative AI is being adopted across your environment? Check out the ready-made dashboard called “Generative AI feature adoption”.
Please note that your prompts are sent to LLMs hosted by enterprise vendors such as Microsoft Azure AI and AWS Bedrock, which power Dynatrace Intelligence generative AI. These vendors don't store the submitted data or any responses. The prompts and related responses are used only to serve your experience. Neither our enterprise vendors nor Dynatrace use the prompts to fine-tune or improve any models or services, or to train models across customers or environments.
Dynatrace may store prompts submitted to Dynatrace Assist and the responses provided by the LLMs to better understand use cases, contextualize feedback on those responses, and identify additional user expectations.
Enable your AI agents with high-quality data and real-time production insights.
Enable your agents with reliable insights, with our toolkit accessible via MCP.
Automate smarter with intelligent detection, forecasting, and custom analysis.
Action based on answers with deterministic insights and agentic operations.
Build agentic workflows to automate, orchestrate and govern operations at scale.
Dynatrace Assist gets a significant upgrade with this release.
The chat interface changes from a modal to a side panel, so you can chat with Assist while keeping your dashboard, notebook, or any other app page in view.
/ in the chat to quickly access and filter all prompt examples.Assist now consults a curated knowledge base of skills and reference files that capture Dynatrace domain expertise. This grounds Assist's reasoning in structured guidance from Dynatrace SMEs, for deeper and more targeted analysis than documentation and the foundation model alone can provide.
The driver LLM behind Assist switches from OpenAI GPT-4o (hosted on Azure) to Anthropic Claude Sonnet 4.6 (hosted on Amazon Bedrock). This means better responses across multi-step reasoning, tool use, and complex investigations.
Our natural-language-to-DQL capability moves from a RAG-based approach to a purpose-built, fine-tuned foundation model based on Llama 3.1 8B, trained on hundreds of thousands of DQL examples. This results in significantly more reliable DQL generation.
Dynatrace Assist now supports agentic capabilities, which let you ask questions and analyze issues from a single conversational interface powered by Dynatrace Intelligence. It combines generative, causal, and predictive AI with environment context (Grail and Smartscape) to deliver transparent, data‑backed recommendations and end‑to‑end remediation guidance.
Read more about the release here, or learn more about how to get started with Agentic mode.
With the launch of Dynatrace Intelligence, the Ask Davis CoPilot app has been renamed to Dynatrace Assist. You'll also see refreshed icons across the interface for clearer, more consistent visuals.
